Hey everyone,
So recently I upgraded my unraid server from Intel i7 4790K system to AMD Ryzen 9 7900X, basically wanted to move back to PC gaming from PS5 after PS5 Pro launch😅. 
Long story short -- ideally I wanted one machine for both Unraid Server task as well as for Gaming but as it would be under VM and my prior experience with 4790k wasn't anything to brag about (mainly low CPU performance which on some level i belived because of weak old hardware) I first wanted to slowly test things before upgrading GPU cause GPU wise even with 4790K system i was able to get bare metal level of performance out of my GTX 1070 TI (only in GPU Heavy Benchmarks).
So fortunately I am getting basically bare metal level of performance of a 7600X from my 7900X (with 6 Cores and 12 Threads) under VM which is what i hoped to achieve. And that too without any XML tweaks on Unraid v7. So now I'm waiting for my RTX 4070 TI Super.
But here's the thing ever since i installed NVME SSD (XPG Gammix S70 Blade) in my server for VM I am getting this error under syslog.
 

Unraid kernel: vfio-pci 0000:0d:00.0: VPD access failed.  This is likely a firmware bug on this device.  Contact the card vendor for a firmware update

I have bound it to VFIO and passed it to my VM but I keep getting this error under my Syslog.

I have already tried checking firmware on it but it already running on latest version. So is it anything to worry or if anyone knows how can I fix this cause its been already over a week and fortunately I am yet to face any issue because it.

So yeah basically wanted to share my success story with those who are under similar dilemma as i once was before finally clicking on buy button and ask around about this error.
If anyone can help me fixing it.... or should I just ignore it....😅
 
 

unraid-diagnostics-20241129-1134.zip

Because you added a pcie device

you had a vfio device previous set to bind for the VM
That older VFIO pci ID has changed with the additon of the nvme (which is a pcie device)

as such you need to remove the vfio bind and reslect the card that was vfio bind.

otehr vm settings may need to be enbaled such as all unasfe intrupts and mutifunciton both

4 hours ago, bmartino1 said:

Because you added a pcie device

you had a vfio device previous set to bind for the VM
That older VFIO pci ID has changed with the additon of the nvme (which is a pcie device)

as such you need to remove the vfio bind and reslect the card that was vfio bind.

otehr vm settings may need to be enbaled such as all unasfe intrupts and mutifunciton both

I was not using VFIO bind prior to adding NVME to my system, the only thing which is VFIO bound is that NVME only which of course i only bounded after adding it to the system😅.

The only other device which i could bind was my GPU but i'm not using VFIO bind with it as i never had any issues relating to gpu utilisation on VM so i'm using my gpu without VFIO bind thing.

 

And as for unsafe interrupts and multifunction on both -- are you sure unsafe interrupts is safe to use cause it's showing that it can cause data corruption on some systems😅.

And just tried booting with multifunction on both, first i still got that warning under syslog and now my gpu and its audio are appearing in separate IOMMU groups so i can't use this anyway.

So i just tried booting without VFIO bind thing and i still got the same error.

 

Unraid kernel: nvme 0000:0d:00.0: VPD access failed.  This is likely a firmware bug on this device.  Contact the card vendor for a firmware update

So seems like firmware issue to me 😓

i'm confused your vfio binding the nvme and using it in the unraid array?

or is this suposed to be a nvme pci passthorugh for a windows vm under unraid?

if so make sure that nvme is in the top slot of mother board. the top nvme is direct conected to the processor and has better support for pcie passthorugh..

This still seems like a broken vfio as your trying to do something to it while it is esental blacklisted...

Edited December 1, 20241 yr by bmartino1
data

go to setting VM:

 

reboot to apply

the go to main flash:
at the bottom in syslinux:

add this data:

 

kernel /bzimage
append initrd=/bzroot intel_iommu=on iommu=pt nvme_core.default_ps_max_latency_us=5500 default_hugepagesz=1G hugepagesz=1G transparent_hugepage=always pci=noaer pcie_aspm=off

* This may fix soem nvme issue you ahve with vfio and setup vm vfio passthrough...

whoudl need you to double check tools > system vfio log...

11 hours ago, Max said:

1I was not using VFIO bind prior to adding NVME to my system, the only thing which is VFIO bound is that NVME only which of course i only bounded after adding it to the system😅.

2The only other device which i could bind was my GPU but i'm not using VFIO bind with it as i never had any issues relating to gpu utilisation on VM so i'm using my gpu without VFIO bind thing.

 

3And as for unsafe interrupts and multifunction on both -- are you sure unsafe interrupts is safe to use cause it's showing that it can cause data corruption on some systems😅.

4And just tried booting with multifunction on both, first i still got that warning under syslog and now my gpu and its audio are appearing in separate IOMMU groups so i can't use this anyway.

1thank you for confimg.

2depends on gcard.

3yes you want both

 

4yes you want that seperate. is a siple xml edit to fix in the vm...